Detection of Changes in the Opening Pressure of Marine Engine Injectors Using Vibration Methods Otkrivanje promjena u početnom tlaku ubrizgača goriva

Injectors of marine engines are, apart from injection pumps, a key element of the fuel injection system. Their proper operation aff ects the quality and time of producing the fuel-air mixture in engine cylinders. The technical condition of the injectors also has an impact on engine performance and the emission of harmful compounds into the atmosphere. One of the basic injector damages occurring during their normal operation is the change of the injector opening pressure. The article presents mathematical relations indicating the occurrence of vibration diagnostic parameters that could potentially determine changes in the opening pressure of individual injectors. A twostage passive-active experiment was carried out during which an injector working with diff erent spring preload force was tested. The presented results of experimental tests confi rm the possibility of determining whether the injector opening pressure is within the permissible range. The presented method does not require interference in the engine operation. measurements can be made during regular operation.


INTRODUCTION / Uvod
Considering the machinery service strategy in a function of its technical condition it is necessary to have a set of reliable diagnostic methods that determine the technical condition of all vital components important for reliability [13]. One of such methods is skilfully used vibration diagnostics. The use of vibration measurements in diagnostic systems allows determining the optimal moment of stopping the machinery to perform necessary repairs. The moment of the appearance of the defect and its propagation to the catastrophic damage can be illustrated using the so-called P-F curve -fi gure 1.
Analysing Figure 1, it should be assumed that the moment of defect does not eliminate the engine from further operation. However, responsible personnel have to be aware of the defect and have tools to monitor its development and assess the impact on the technical condition of the engine. The main objective of the injection installation is to prepare appropriate doses of fuel with their proper spraying in the cylinder at a certain crankshaft angle before TDC. Each of the mentioned parameters describing the operation of the injection installation is characterized by specifi c limit values given each time by the engine producers. As long as these values do not exceed the permissible ones, the engine is in a state of technical effi ciency. After exceeding the limit values, the object changes its state to the state of technical disability, however, it is still able to be operated. It is crucial that the technical service personnel are aware of the need to quickly repair the resulting and developing defect. If an existing defect is not detected at this stage, it usually develops very quickly leading to severe engine damage.
Based on many years of diagnostic experience gathered at the Institute of Ship Design and Exploitation of Polish Naval Academy, analysis of the occurrence of operational incapacity of marine engines operated in the Navy was carried out - Figure  2 [10]. The conducted analysis showed that the most common defects occur in the following functional systems: -engine fuel supply system -72%, -timing system -19%, -engine power supply system -9%. Damage to injection systems is 49.68% of all damages that occurred in the operation of marine engines in Polish Navy in recent years, of which 51% was damage to the injectors [11].
Exceeding the permissible values characterizing the operation of injectors of marine diesel engines has from the diagnostic point of view two consequences, i.e. leads to an increase or decrease in combustion pressure and changes its distribution as a function of crankshaft angle ( o CA). Of course, there are also many other phenomena that are unfavourable to the engine, but they are not the subject of this paper. Changes in an injectors technical condition cause not only a decrease in the quality of cylinder processes but also in the long-term. They can cause severe secondary damage to other components of the naval propulsion system (e.g. as a result of drastic lowering of the injector opening pressure poorly sprayed fuel can mix with the oil and lead to seizing of the bearings).
The exploitation of the marine engine fuel supply system takes place in marine conditions, which are considered the most diffi cult, and it is refl ected in the high intensity of damage [1,8]. This is due to the destructive impact and use of the engine at low and variable loads, the possibility of water in the fuel, as well as salt and seawater in the intake air, cyclical overloads, etc. The operation of the engine with injectors in a not perfect condition always leads to a decrease in the performance and effi ciency of the marine propulsion system. The reasons for this condition can be varied [12]: -bunkering of low-quality fuel (low quality of bunkered fuel, ineff ective operation of centrifuges, contaminated fi lters, contaminated tanks, leaks of other substances to the fuel supply system, etc.), -lack of immediate reaction of technical personnel to minor primary damage of structural elements, aff ecting the quality of the fuel system operation, -repeated engine starts, frequent changes of its load and longlasting work at low loads, -chemical corrosion caused by the aggressive infl uence of the factors contained in fuel (sulphur and vanadium), especially during long engine stoppages, -erosive eff ect of solid particles contained in the fuel, as well as cavitation erosion of fl ow channels, -thermal eff ect of the heated engine on precise dosing elements. Many scientists devote their attention to issues related to the diagnostics of diesel engine injectors [1,2,3,4,5,6,7,8,9,12,14,15,16,17]. Currently, the most eff ective way to assess the technical condition of injectors is their disassembly and testing on the dedicated test benches [14]. However, this means the engine has to be stopped for the duration of the test. In addition, engine manufacturers recommend testing during reviews carried out at specifi c intervals of time, unfortunately, the damage is very common in the inter-review period.

INJECTOR OPERATION MATHEMATICAL MODEL / Matematički model rada ubrizgača
Before the research, a thorough mathematical analysis of dynamic interactions occurring during the operation of the injector was carried out. Fuel for the injector ( Figure 3 shows the injector of the Sulzer 6 AL. 20/24 engine, which was the object of tests), is supplied by high-pressure pipes in which high pressure is generated by the injection pump. Due to the small volume of the injector chamber and the use of a non-return valve to prevent a pressure drop in the line between successive piston stroke of the pump, high pressure can be obtained very quickly. The injector spring pushes on the needle. Injection starts when the force acting on the needle caused by the fuel pressure exceeds the force generated by the initial spring tension. Injection stops when the spring tension force is suffi cient for the injector to close rapidly. The force of the needle's impact on the seat is directly proportional to the elastic force of the spring. Also, there is a  In the physical model, it was assumed that the system had one degree of freedom, connected with the direction of the injector needle movement. The fl ow at the injector's nozzle is laminar. The spire and the spindle were included as one mass -m.

Figure 4 Kinematic scheme of a precise injector pair Slika 4. Kinematska shema preciznog para ubrizgača
The injector ( Fig. 4) consists of an outer cylinder, a piston of mass m, and a nozzle with holes of diameter d (in the case of the 6 AL 20/24 engine the nozzle has 7 holes). The injector chamber is fi lled with a viscous fuel with dynamic viscosity μ [Ns/m 2 ]. The fi rst task is to formulate the polar equation of the injector. It is the relationship between the force acting on the piston and its speed in relation to the housing of the injector. Due to the occurrence of damping force in the model, the law of Hagen-Poiseuille was used. It assumes that at a stationary, laminar fl ow of incompressible, viscous fl uid in a cylindrical pipe (i.e. in a pipe with a constant, circular cross-section), fl ow volume is proportional to the pressure gradient along the duct, and therefore to the pressure diff erence at the ends of the duct (1) [10]: (1) where: Δp -pressure drop over the length of one-meter pipe of diameter d, μ -dynamic viscosity of fuel, Q -fuel fl ow rate in the pipe, which might be described as: (2) where: u -fuel speed in the pipe Speed is determined by the law of fl ow continuity, with: (3) where: -number of nozzle holes, -diameter of nozzle hole.
It has resulted: (4) Assuming that the diameters of the injector holes are much smaller than the diameter of piston D, force F acting on the piston reaches the value of: (5) By substituting the formula for velocity u, the polar equation of fuel's damping was obtained in the formula: (6) where the coeffi cient of viscous damping is defi ned as: (7) The model of needle impact into the cylindrical body of the injector was simplifi ed to the form of a vibration forced by instantaneous force. The measure of the short-term impact of a needle is the impulse S, i.e.: (8) Because the pulse duration is insignifi cantly short, i.e.
(9) A simplifying assumption about the occurrence of instantaneous force was made. If the value of force F 0 is constant then the condition is: (10) It is a unit impulse. Determining the impulse response of the system with the damping of fuel inside the injector chamber in oscillating systems with subcritical damping allows recording the motion of the piston in the form of: (11) where: -constant, -initial phase.
The equation of motion takes the form: (12) Applying the theorem on momentum change: (13) Any fi nite impulse acting on a body with mass m will result in a fi nite increase in the speed of its body. If the body had an initial speed of 0 at the end of the pulse will be: (14) When estimating the impulses of and forces that act on the injector in time Δt, it should be noted that the mass of the plunger moving during the infi nitely small Δt changes its position during an end impact of Δx in the time of Δt, hence: 15) where: 0(ε) is the small size of the fi rst order.
This means that the pulses from the injector spring and the fuel damping are the fi rst and the second-order impulses and do not aff ect the mass speed after the pulse. Thus, the initial conditions are: (16) Substituting it result: (17) Which could be defi ned as: (18) where: Finally: The obtained equation (20) indicates that the change of the value of the force acting on a needle connected with the injector pre-tension spring changes the form of the oscillation equation for the needle. This means that there is a vibration parameter that identifi es the injector adjustment status.

COURSE OF RESEARCH / Tijek istraživanja
As a form of experiment realization, a two-stage passiveactive experiment was adopted, with the simulated variable state which was the injector opening pressure. Along with the change of the fuel injector opening pressure, the fuel injection angle was also changed indirectly, and thus the moment and value of the maximum combustion pressure. During the fi rst stage of the tests, characteristics of the injector opening pressure as a function of the position of the adjusting screw were made. It was established that change of the position of the adjusting screw by 90 o causing the reduction of the spring preload and results in a change of injector opening pressure by 3 MPa. During the adjustment of injector on the test stand, the time waveforms of vibration accelerations were recorded at two points. The fi rst one was located on adjusting screw (signal2). Adjusting screw in this injector type is equipped with an internally screwed hole, thanks to which it was possible to screw in a threaded rod, the length of which was chosen so that after mounting the injector it was possible to discharge its upper part through the hole in the engine's head cover. As a result, it was possible to measure the vibrations of the injector which is not available during normal engine operation. The second measuring point was located on the injector high-pressure pipe (signal3). The tested injector during the regulation at the testbed is shown in Figure 4. Measurements of vibration parameters were carried out for the nominal injector opening pressure of 24,5 MPa and lower pressures (21,5; 18,5; 15,5; 12,5; 9,5; 6,5; 3,5 MPa). The vibration parameters for pressures above the nominal pressure have not been verifi ed because situations of increase the injector opening pressure during its normal operation do not occur (they can only result from incorrect adjustment).
During the second stage of tests, the injector with known characteristics of pressure changes as a function of the initial tension of the spring was mounted on the engine. Measurements of vibration accelerations were carried out in three points -two identical as during the fi rst stage and third located on the engine head tighten bolt (signal4). The engine torque receiver was a water brake. Tests were carried out for three engine crankshaft's speeds and loads with a torque due to the marine engine propeller characteristics: 1. n = 500 rpm, M = 2050 Nm; 2. n = 600 rpm, M = 3420 Nm; 3. n = 700 rpm, M = 4390 Nm.
For all engine speeds, the spring preload was changed according to the characteristics determined during the fi rst test stage. The technical solution used allowed to change the opening pressure of the injector without stopping the engine. This procedure made it possible to carry out measurements for diff erent injector opening pressures under the same engine operating conditions. The authors did not meet in the literature description of an implementation of injector tests in a similar way. Usually the researchers, in all the changes in the spring characteristics of the injector, disassemble it from the engine. Such operation makes the engine conditions impossible to be reproduced during the next steps of measurements (the exception is the use of an electro-eddy rotor brake). To determine the eff ect of changes in the injector opening pressure on the indicated pressure parameters, continuous measurement of this pressure was carried out. The obtained results of changes in the indicated pressure at the rotational speed of 700 rpm are shown in Figure 5.   In the case of higher rotational speeds (600 and 700 rpm) the maximum pressure spread is even smaller and does not exceed 5%. These are the quantities that generally cannot raise the crew's concern. What is more, on the indicator diagrams there was also no change in the crank angle occurrence of the maximum combustion pressure depending on injector opening varying pressure. According to the above, detection of injectors damage connected with changes of spring tension using the indication of the engine may be impossible regardless of the value changes value of this tension.

CHOSEN TEST RESULTS / Rezultati odabranoga testa
Vibration accelerations time courses recorded during both stages of the tests were subjected to a two-stage analysis. First, the overall levels of vibration accelerations recorded on injector test and regulation station were compared. Signals were averaged in two-second periods. The waveforms recorded on engine test bed during its operation were averaged over the time corresponding to the full cycle of its operation (two revolutions, this time was diff erent for each rotational speeds). An example of the obtained results for the measurement carried out on the high-pressure injector pipe, carried out at the injector test and regulation station is shown in Figure 7.
Analysis of Figure 7 allows indicating the relationship between the tested vibration parameters and the injector's opening pressure. The higher the opening pressure, the higher the vibration acceleration values. This is unusual situation for vibration diagnostics, where diagnostics have become accustomed to a situation in which increasing vibration parameters indicate a deterioration of the technical condition of the machine or its element. This phenomenon can be explained easily, namely that the higher the injector opening pressure, the greater the dynamics of the injection process, causing the creation of force with higher values. The next stage of the analysis was an analysis of vibration accelerations amplitude spectra in the full range of measured frequenciesup to 25,6 kHz. An example of spectra obtained from the time courses recorded at the injector test stand is shown in Figure  8. These spectra were then examined for the determination of the frequency band sensitive to changes in the injector opening pressure values. A range of diff erent frequency bands was used, from which the average amplitude of vibration accelerations was calculated. As a result of the analysis, the frequency band in the range of 2000-2800 Hz was selected as the most sensitive. Figure 7 Overall values of vibration acceleration a rms for two-second time periods counted for the 1Hz-10kHz frequency band and all tested injector opening pressures for measurements carried out at the high-pressure injector pipe  Table 1 summarizes the results of the average values of vibration accelerations (a rms ) obtained from spectra calculated of signals recorded on the injector control bolt (signal 2 -sig. 2) and the high-pressure connector pipe (signal 3 -sig. 3).
An analysis of values presented in table 1 allows to state that the chosen frequency range is a band sensitive to changes in vibrations accelerations amplitude associated with decreasing the opening pressure of the investigated injector. A signifi cant reduction in the average amplitude values calculated in the indicated frequency range occurs when the injector opening pressure decreases by at least 12 MPa. In the same way, the waveforms recorded during engine operation on test stand were analysed. An example of the results obtained with reference to the average values of vibration accelerations calculated from time courses was shown in Figure 9. The best results in terms of average values of vibration accelerations calculated from time course fragments were obtained for the rotational speed of 700 rpm and the measuring point located on the high-pressure connector pipe. In the case of other rotational speeds and other measuring points, the results are ambiguous. Therefore, the spectral analysis of all vibration acceleration recorded during the operation was carried out. The example was shown in Figure 10. The analysis of spectra presented in Figure 10 draws attention to their utterly diff erent course than in the case of spectra obtained during measurements on injectors testing and adjustment station. There are two main reasons for this. The fi rst of them is an entirely diff erent mounting of the injector on the engine and thus a change in the stiff ness of the system. The second is the interaction of other working engine components, including the remaining injectors  working with the same frequency. Despite the occurrence of these phenomena signifi cantly changing the obtained vibration acceleration spectra, the frequency range of 2-2,8 kHz remains sensitive to changes in the injector opening pressure. This is evidenced by the results of all measurements, which were collected in Table 2. To facilitate the analysis of the obtained results, they were also presented graphically on fi gures 11 -13. Analysis of the presented results allows to state that the most unambiguous dependencies occur for measurements conducted at engine rotational speed of 700 rpm. This situation is repeated for all measuring points, however, the most visible dependencies are observable for the measuring point located on the timing cover bolt. The signals obtained for the rotational speed of 500 rpm are ambiguous and cannot be used as a diagnostic material due to the low sensitivity to the propagation of simulated damage. Slight changes in the opening pressure (which may be, for example, due to the natural "subsidence" of the spring) are impossible to detect. They even cause an increase in the value of the vibration parameter being studied, where its further propagation causes its reduction.

CONCLUSIONS / Zaključci
The conducted tests and its results confi rm the eff ectiveness of vibration diagnostics and the ability to implement the analysis of average values of vibration accelerations from selected frequency ranges to detect ineffi ciencies associated with the change (reduction) of the injector opening pressure in marine diesel engines.
The possibility of detection of changes in the opening pressure occurs when the value of the opening pressure changes by at least 50% to the nominal pressure. This value may seem to be large, however, vibration acceleration as a diagnostic parameter is characterized by a higher sensitivity than indicated pressure measurement, especially at higher rotational speeds. The equation 20 presented in the paper is correct, which is confi rmed by the results collected in Table 1, but it does not take into account the infl uence of other engine components, but only the changes within the hull of the tested injector. It should be noted that marine diesel engine is a challenging object from a diagnostic point of view, mainly due to the high complexity of the structure. The extortion in multi-cylinder engines occurring at the same frequency is also a critical diffi culty. A solution might be the analysis of vibration signals in the function of crankshaft rotation, which the authors plan to do in the near future.